Biochemistry 2005, 44, 7583-7592
7583
Development of Small Designer Aldolase Enzymes: Catalytic Activity, Folding,
and Substrate Specificity†
Fujie Tanaka,* Roberta Fuller, and Carlos F. Barbas, III*
The Skaggs Institute for Chemical Biology and Departments of Chemistry and Molecular Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
ReceiVed February 4, 2005; ReVised Manuscript ReceiVed March 22, 2005
ABSTRACT: Small (24-35 amino acid residues) peptides that catalyze carbon-carbon bond transformations
including aldol, retro-aldol, and Michael reactions in aqueous buffer via an enamine mechanism have
been developed. Peptide phage libraries were created by appending six randomized amino acid residues
to the C-terminus or to the N-terminus of an 18-mer R-helix peptide containing lysine residues. Reactionbased selection with 1,3-diketones was performed to trap the amino groups of reactive lysine residues
that were necessary for the catalysis via an enamine mechanism by formation of stable enaminones. The
selected 24-mer peptides catalyzed the reactions with improved activities. The improved activities were
correlated with improved folded states of the peptides. The catalyst was then improved with respect to
substrate specificity by appending a phage display-derived substrate-binding module. The resulting 35mer peptide functioned with a significant proportion of the catalytic proficiency of larger protein catalysts.
These results indicate that small designer enzymes with good rate acceleration and excellent substrate
specificity can be created by combination of design and reaction-based selection from libraries.
Generation of enzyme-like small designer peptide catalysts
that achieve the proficiency of natural enzymes with desired
substrate specificity is a challenging task. Designer protein
catalysts have been prepared by engineering naturally occurring enzymes (1-4) and by using antibody libraries (i.e.,
immune diversity) and selection with designed small molecule compounds (5-7). For experimental purposes, small
peptides are more useful and convenient than large proteins,
as long as they provide the same function, because small
peptides can be more easily prepared and their characterization is simpler. However, fundamental questions remain to
be answered: Can small peptides attain the catalytic efficiency of larger protein catalysts, and how can these peptide
catalysts best be developed? The folded states of enzymes
are key to their catalytic activity since structure can be used
productively to modify the chemical reactivity of amino acid
side chains. Small peptides are often limited in their potential
to catalyze reactions because of the limited ability to adopt
well-defined structures. Small peptides have also demonstrated limited specificity for small substrate molecules.
For small catalytic peptides (<50 amino acid residues)
that function in aqueous buffer, rational design has been used
for the creation of catalysts of the decarboxylation of
oxaloacetic acid (8-12), ester hydrolysis (13-15), and
transesterifications (13). Rational design, however, is limited
in throughput. A combination of rational design and librarybased methods with reaction-based selection, which has been
used for the development of antibody catalysts (7), may be
†
This study was supported in part by the NIH (Grant CA27489)
and The Skaggs Institute for Chemical Biology.
* To whom correspondence should be addressed. C.F.B.: e-mail,
carlos@scripps.edu; fax, 858-784-2559; phone, 858-784-9098. F.T.:
e-mail, ftanaka@scripps.edu; fax, 858-784-2559; phone, 858-784-2559.
an attractive approach to small peptide catalyst development
(16, 17). Here we have explored strategies for the creation
of small designer enzymes. We have developed small
peptides that catalyze carbon-carbon bond formation and
cleavage via an enamine mechanism. First, reaction-based
selection from peptide phage libraries was performed to
develop peptides that possessed improved catalytic activity.
The peptide was then improved with respect to specificity
for a small molecule substrate by appending a phage displayderived substrate-binding domain. To understand key factors
critical for catalysis performed by small peptides, we have
characterized the peptides with respect to catalysis, folded
state, and substrate specificity. This research addresses how
small peptides with improved catalytic activity and with
improved substrate specificity can be created and how the
catalytic activity of small peptides might approach that of
larger protein catalysts.
MATERIALS AND METHODS
Construction of the XXXXXXYKLLKELLAKLKWLLRKL
Library [(NNK)6-YLK Library]. Oligonucleotides YLK18opt-SfiI-HindIII-f (5′-CGGCCTACAAGCTTCTGAAGGAACTGCTCGCTAAACTGAAGTGGCTGCTCCGTAAACTGGGCCAGGC-3′) (10 µg) and YLK18opt-SfiI-HindIII-b (5′-TGGCCCAGTTTACGGAGCAGCCACTTCAGTTTAGCGAGCAGTTCCTTCAGAAGCTTGTAGGCCGCCT-3′) (10 µg) were mixed in a
total volume of 40 µL, and the mixture was heated at 95 °C
for 2 min and then cooled to 25 °C. This mixture was ligated
to SfiI-digested pComb3X to give construct A. Oligonucleotides YLK-18opt-SfiI-f (5′-CGGCCTACAAACTGCTGAAGGAACTGCTCGCTAAACTGAAGTGGCTGCTCCGTAAACTGGGCCAGGC-3′) (10 µg) and
10.1021/bi050216j CCC: $30.25 © 2005 American Chemical Society
Published on Web 04/26/2005
7584 Biochemistry, Vol. 44, No. 20, 2005
YLK-18opt-SfiI-b (5′-TGGCCCAGTTTACGGAGCAGCCACTTCAGTTTAGCGAGCAGTTCCTTCAGCAGTTTGTAGGCCGCCT-3′) (10 µg) were mixed in a total volume
of 40 µL, and the mixture was heated at 95 °C for 2 min
and then cooled to 25 °C. This mixture was ligated to SfiIdigested pComb3X to give construct B. Construct B was
used as a template for PCR using primers EcoRI-f (5′TTTTGAATTCAGGAGGAATTTAAAATGAAAAAGACAGGAACTATCGCGATTGCAGTGGCACTG-3′) and HindIII-NNK6-SfiI-b (5′GAGGAGGA-GAAGCTTGTA-(MNN) 6 GGCCGCCTGGGCCACGGT-3′), where M ) A or C and N ) A, C, G,
or T. The PCR conditions were as follows: 1 µg of each
primer, 100 ng of the template, 2.5 mM each dNTP, and 2.5
units of Taq polymerase in a total volume of 100 µL, with
a program of 94 °C for 30 s; 94 °C for 15 s, 56 °C for 15 s,
72 °C for 1 min (25 times); and 72 °C for 5 min. The PCR
product was digested with EcoRI-HindIII, and the digested
fragment was ligated to EcoRI-HindIII-digested construct
A to give the (NNK)6-YLK library in pComb3X.
Construction of the YKLLKELLAKLKWLLRKLXXXXXX
Library [YLK-(NNK)6 Library]. Oligonucleotides YLK18-f
(5′-GAGGAGGTGGCCCAGGCGGCCTACAAGCTTCTGAAGGAACTGCTCGCTAAACTGAAGTGGCTGCTCCGTAAACTG-3′) (10 µg) and YLK18-NNK6-b [5′GAGGAGGAGGCCGGCCTGGCC(MNN)6CAGTTTACGGAGCAGCCA-3′] (10 µg) and 2.5 mM each dNTP were
mixed with 10× NEBuffer 1 (New England Biolabs, NEB)
(10 µL) and with 10× NEBuffer 2 (NEB) (10 µL) in a total
volume of 196 µL, and the mixture was heated at 95 °C for
2 min. After being cooled to 25 °C, 20 units of DNA
polymerase I, large (Klenow) (5 units/µL, 4 µL), was added
to the mixture. After 15 min at 25 °C, the enzyme was
denatured at 75 °C for 20 min. The mixture was EtOH
precipitated and digested with SfiI followed by purification
using Centri-spin-20 (Princeton Separation). The SfiI-digested
fragment was ligated to an SfiI-digested pComb3X to give
the YLK-(NNK)6 library in pComb3X.
Binding Selection against Diketone Haptens Using a
Phage Display System. The ligation mixture was ethanol
precipitated and transformed into E. coli ER 2537 cells by
electroporation (three transformations per library). After
transformation, SOC (2% tryptone, 0.5% yeast extract, 0.05%
NaCl, 2.5 mM KCl, 10 mM MgCl2, 20 mM glucose) (5 mL)
was immediately added into the each reaction, and the cells
were grown at 37 °C. After 1 h, super broth (3% trypton,
2% yeast extract, 1% Mops) (SB) (10 mL) and carbenicillin
(50 µg/mL) were added. The library size was determined to
be 4.7 × 107 for the (NNK)6-YLK library and 4.6 × 107
for the YLK-(NNK)6 library, respectively, by plating of this
stage culture (10, 1.0, and 0.1 µL) onto carbenicillin (100
µg/mL) plates of LB agar (carb plates). After 2 h, helper
phage VCS-M13 (7.8 × 1012) were added, and the three
cultures for each library were combined into SB (total volume
150 mL) containing carbenicillin (50 µg/mL). After 2 h,
kanamycin (70 µg/mL) was added, and the culture was
incubated at 37 °C overnight. The cells were removed by
centrifugation (4000 rpm, 30 min), to the supernatant was
added poly(ethylene glycol) 8000 [final concentration 4%
(w/v)] and NaCl [final concentration 3% (w/v)], and after
30 min on ice, the mixture was centrifuged (9000 rpm, 20
min) (PEG precipitation). The phage precipitate was resus-
Tanaka et al.
pended in 1% BSA1/PBS (10 mM Na2HPO4, 1.8 mM KH2PO4, 137 mM NaCl, 2.7 mM KCl, pH 7.4) (5.0 mL) and
filtered (0.2 µm).
Wells of a microtiter plate (Costar 3690) were coated with
1-BSA (two wells per library) and 2-BSA (two wells per
library) (3-5 µg/25 µL of PBS per well) at 4 °C overnight,
washed with H2O two times, and blocked with 3% BSA/
PBS (170 µL per well) at room temperature for 1.5 h.
Blocking solution was removed, and the library phage (25
µL per well) was added. After 2 h, the wells were washed
10 times with 0.5% Tween 20/PBS (PBST) to remove
unbound phage, and then the bound phage were eluted by
trypsin digestion (Difco trypsin 1:250, 10 mg/mL in TBS,
50 µL per well) at 37 °C for 30 min. The elution was added
to Escherichia coli ER2537 cells (15 mL in SB). The eluted
phage were combined on the basis of the diketones. The
culture was incubated at 37 °C. After 15 min, the infected
cell culture (30 µL) was diluted with SB (270 µL) and plated
(1, 10, and 100 µL) on carb plates to determine the output.
Carbenicillin (50 µg/mL) was added to the infected cell
culture and shaken at 37 °C for 1 h. Helper phage VCSM13 (1 mL, 2.6 × 1012 pfu) were added, and the culture
was diluted into 100 mL with SB containing carbenicillin
(50 µg/mL). After 2 h, kanamycin (70 µg/mL) was added,
and the culture was incubated at 37 °C overnight. After
centrifugation and PEG precipitation as described above, the
phage precipitate was resuspended in 1% BSA/PBS (pH 7.4)
(2.0 mL) and filtered (0.2 µm). This phage solution was used
for further panning. For the second and third rounds, phage
were added to the BSA-coated wells, incubated for 30 min
at room temperature to exclude the phage that bound to BSA,
and then transferred to hapten-coated wells.
Screening Functional Peptides and Sequence Analysis.
After the third round panning, the eluted phage were used
to infect E. coli ER2537 cells, and the peptide-pIII fusion
was used for the ELISA to screen for the clones that
produced soluble peptides that bound to the diketones.
Individual colonies were picked from the plates and grown
in SB (5 mL) containing carbenicillin (50 µg/mL) at 37 °C
for 7 h. Expression of the Fab genes was induced by addition
of IPTG (1 mM), and the cultures were incubated overnight.
The culture was centrifuged (3500 rpm, 20 min), and the
supernatant was used for assessing the binding activity by
ELISA using 1- and 2-BSA. For the screening of soluble
peptides (peptide-decapeptide fusions) that bind to the
diketones, 20 clones for the library panned with 1-BSA and
10 clones for the library panned with 2-BSA were assayed.
FT-YLK3 and FT-YLK4 were found in the library panned
with 2-BSA. For the screening of peptide-pIII fusions that
bound to the diketones, 20 clones for the library panned with
1-BSA and 20 clones for library panned with 2-BSA were
assayed. FT-YLK25 was identified from ELISA of peptidepIII fusion.
The ELISA was performed using the following procedure.
Microtiter plates (Costar 3690) were coated with 1-BSA
(0.5 µg/25 µL of PBS per well) at room temperature
overnight, washed two times with H2O, and blocked with
3% BSA/PBS (50 µL per well) at room temperature for 1 h.
Blocking solution was removed, the ELISA sample (25 µL
1 Abbreviations: BSA, bovine serum albumin; ELISA, enzymelinked immunosorbent assay; ee, enantiomeric excess.
Small Designer Aldolase Enzymes
per well) was added, and the plate was incubated at room
temperature for 2 h. The well was washed 10 times with
H2O. The bound peptide was detected using alkaline phosphatase conjugated anti-decapeptide antibody and the phosphatase substrate p-nitrophenyl phosphate. The resulting
yellow color was measured at 405 nm.
The nucleotide sequence of the peptide encoding regions
was analyzed using 5′-primer ompseq (5′-AAGACAGCTATCGCGATTGCA-3′).
Peptides. Peptides were synthesized on a peptide synthesizer using standard Fmoc solid-phase peptide synthesis
chemistry, were purified by HPLC to >95% purity, and were
characterized by electrospray mass spectrometry.
Enaminone Formation (Figure 1). Reactions were initiated
by adding 5 µL of 2,4-pentanedione (100 mM in CH3CN)
to 95 µL of peptide solution in a 96-well UV plate (Costar
3635) at 25 °C. The final conditions were [peptide] 100 µM
and [2,4-pentanedione] 5 mM in 5% CH3CN-42.5 mM
sodium phosphate buffer (total volume 100 µL). The
enaminone formation was measured by following the difference in absorption at 318 nm and at 400 nm for 2 h 40
min using a microplate spectrophotometer.
Catalytic Assay and Kinetics for Retro-Aldol Reactions
of 4 and 5. Reactions were initiated by adding 5 µL of a
stock solution of substrate 4 or anti-5 in CH3CN to 95 µL
of peptide solution in sodium phosphate buffer (pH 7.5) at
25 °C for a 96-well plate (Costar 3915). The final conditions
were [peptide] 100 µM and [substrate] 250 µM-2 mM in
5% CH3CN-42.5 mM sodium phosphate buffer (pH 7.5)
(total volume 100 µL) except where noted. The increase in
the fluorescence (λex 330 nm, λem 452 nm) was monitored
by using a microplate spectrophotometer. The observed rate
was corrected for the uncatalyzed rate in the absence of
peptide. The kinetic parameters, kcat and Km, were determined
by curve fitting of S-V plots as described by the MichaelisMenten equation using MacCurveFit (Kevin Raner).
Catalytic Assay and Kinetics for Retro-Aldol Reactions
of 12. The reaction was initiated by adding 8 µL of a stock
solution of 12 in CH3CN to a mixture of 4 µL of 1 mM
peptide in H2O and 68 µL of 50 mM sodium phosphate (pH
7.5) at 25 °C. The formation of the product aldehyde was
measured by HPLC with 10 µL injections of the reaction
mixture. Analytical HPLC was performed using a MicrosorbMV C18 analytical column (Varian) eluted with 40% CH3CN-0.1% aqueous trifluoroacetic acid (TFA) at a flow rate
of 1.0 mL/min, with detection at 254 nm. The retention time
of the product aldehyde was 11 min, and that of 12 was 7.0
min. The initial velocities were determined from the linear
range of the rate. Since nearly equal peptide-substrate
concentrations were employed in the assay, the kcat and Km
were determined according to the procedure reported by
Smith et al. (18).
Retro-Aldol Reactions of D-Erythrose. Reactions were
initiated by adding 8 µL of a stock solution (50 mM) of
D-erythrose in H2O to 72 µL of peptide solution in sodium
phosphate buffer (pH 7.5) at 25 °C. The final conditions were
[peptide] 50 µM and [D-erythrose] 5 mM in 42.5 mM sodium
phosphate buffer (pH 7.5) (total volume 80 µL). To follow
the reaction, an aliquot of the reaction (12 µL) was mixed
with 1 µL of 250 mM p-nitroaniline in MeOH, 1 µL of CH3COOH, and 1 µL of 2 M NaBH3CN in H2O, and after 10
min, the mixture (10 µL) was injected onto HPLC to analyze
Biochemistry, Vol. 44, No. 20, 2005 7585
the product 2-[(4-nitrophenyl)amino]ethanol. The analytical
HPLC was performed using a Microsorb-MV C18 analytical
column at a flow rate of 1.0 mL/min, with detection at 400
nm. The elution was 8% CH3CN-0.1% aqueous TFA, and
the retention time of 2-[(4-nitrophenyl)amino]ethanol was
17.1 min.
Catalytic Assay for Aldol and Michael Reactions. Reactions were initiated by adding 5 µL of a stock solution of
substrate aldehyde in CH3CN to a mixture (95 µL) of acetone
and of peptide solution at 25 °C. The rate of the reaction
was measured by HPLC detection of 4, 9, or 11 with a 10
µL injection of the reaction mixture. The analytical HPLC
was performed using a Microsorb-MV C18 analytical column
at a flow rate of 1.0 mL/min, with detection at 254 nm. For
the aldol reaction to afford 4, the elution was 40% CH3CN-0.1% aqueous TFA, and the retention times of 4 and 7
were 5.6 and 9.6 min, respectively. For the aldol reaction to
afford 9, the elution was 30% CH3CN-0.1% aqueous TFA,
and the retention times of 9 and 8 were 5.4 and 7.0 min,
respectively. For the Michael reaction to afford 11, the
elution was 15% CH3CN-0.1% aqueous TFA, and the
retention times of 11 and 10 were 13.4 and 16.3 min,
respectively. The enantiomeric excess of 9 was determined
by HPLC analysis using a Chiralpak AD-RH (Daicel)
column. The elution was 30% CH3CN-0.1% aqueous TFA,
and the retention times of (R)- and (S)-9 were 15.0 and 17.9
min, respectively, at a flow rate of 1.0 mL/min with detection
at 254 nm.
RESULTS AND DISCUSSION
Mechanism-Based Selection of Peptides. Many carboncarbon bond formation and cleavage reactions in biological
systems proceed through an enamine mechanism (19-22).
Enzymes often use lysine -amino groups to form an enamine
in their catalytic reactions; however, most -amino groups
of lysine residues are not nucleophiles. Special environments
within the three-dimensional folded structures of enzymes
are required in order for the amino group to act as a
nucleophile for enamine formation with carbonyl compounds.
To access peptides possessing such reactive residues, reaction-based selection using combinatorial libraries combined
with peptide design should be useful. We have recently
demonstrated that reaction-based selection using phage
display with 1,3-diketones can be used to access antibody
catalysts operating via an enamine mechanism (23, 24). The
diketones were previously used to modify nucleophilic
-amino groups of lysine residues of enzymes by the
formation of enaminones (25, 26). In addition, antibodies
that catalyze aldol reactions (aldolase antibodies) via an
enamine mechanism were also generated by reactive immunization with 1,3-diketone derivatives (7, 27). Therefore,
to develop small peptide catalysts that operate via an enamine
mechanism, we have performed reaction-based selection with
1,3-diketones (28).
Amine-catalyzed decarboxylation and aldol reactions share
imine and enamine intermediates on their reaction coordinates. Accordingly, enzyme and antibody catalysts of these
reactions are generally bifunctional catalysts capable of
accelerating both decarboxylation and aldol reactions (29,
30). To select peptides that operate via an imine/enamine
reaction mechanism, we sought to exploit mechanistic
7586 Biochemistry, Vol. 44, No. 20, 2005
Tanaka et al.
Scheme 1: Diketones, Enaminone Formation, Modification with N-Acyl-β-lactam, and Retro-Aldol Reactions
Table 1: Sequences of Peptides
peptide name
sequence (C-terminal amide)
YLK-18opt
FT-YLK3
FT-YLK4
FT-YLK25
FT-YLK3-R5
FT-YLK3-NAc
FT-YLK3-23S
FluS303-FTYLK3
YKLLKELLAKLKWLLRKL-NH2
YKLLKELLAKLKWLLRKLLGPTCL-NH2
YKLLKELLAKLKWLLRKLSDHLCL-NH2
SPFLGQYKLLKELLAKLKWLLRKL-NH2
YRLLRELLARLRWLLRRLLGPTCL-NH2
Ac-YKLLKELLAKLKWLLRKLLGPTCL-NH2
YKLLKELLAKLKWLLRKLLGPTSL-NH2
YPNEFDWWDYYYKLLKELLAKLKWLLRKLLGPTCL-NH2
homology by using as our starting point an 18-residue peptide
(YLK-18opt) (9) that catalyzes the decarboxylation of
oxalacetic acid. This peptide contains multiple lysine residues
that are critical for its activity. To evolve the catalytic activity
of this peptide, a six amino acid residue library was appended
to the C-terminus or to the N-terminus of the peptide, and
the libraries YKLLKELLAKLKWLLRKLXXXXXX and
XXXXXXYKLLKELLAKLKWLLRKL (X ) any of the
natural 20 amino acids) were displayed on phage. Reactionbased selection with 1,3-diketones 1-BSA (bovine serum
albumin) and 2-BSA was performed from the peptide phage
libraries (31) to trap the amino groups of reactive lysine
residues that were necessary for the catalysis via an enamine
mechanism by formation of stable enaminones (Scheme 1).
Enzyme-linked immunosorbent assay (ELISA) was used
to identify the clones that bound to 1- and 2-BSA from
the selected libraries. The three peptides that gave the
strongest ELISA signals, FT-YLK3, FT-YLK4, and FTYLK25, and the parental peptide YLK-18opt were chemically synthesized (C-terminal amide). Peptide sequences are
shown in Table 1. Since the solubility of peptide FT-YLK4
was low in aqueous buffer at neutral pH, only FT-YLK3
and FT-YLK25 were characterized.
Enaminone Formation and Reaction with an N-Acyl-βlactam. The peptides were expected to bind diketones through
enaminone formation rather than through noncovalent interactions. The ability of the peptides to form the enaminone
was analyzed using 2,4-pentanedione, the smallest compound
containing a 1,3-diketone structure. Peptides FT-YLK3, FTYLK25, and YLK-18opt formed a UV-observable enaminone
as monitored at 318 nm (23-28). The velocities of the
enaminone formation of FT-YLK3 and of FT-YLK25 with
2,4-pentanedione were faster than that of YLK-18opt at
neutral pH. Selection with diketones afforded peptides with
improved rates of enaminone formation compared to the
parental peptide. The pH profile of the velocity of the
enaminone formation indicated that FT-YLK3 had a pKa
lower than that typical of a lysine -amino group (pKa 10.67)
(32) or an N-terminal amino group, and the lowest pKa was
approximately 5.5 (Figure 1). The selected peptides had
neutral amino group(s) at neutral pH that could serve as the
nucleopiles for formation of enaminone with the diketones
and for catalysis of enamine-based reactions. To provide
information about how many of the amino groups of FTYLK3 were able to form enaminone, i.e., the number of
nucleophilic amino groups, the peptide was treated with an
N-acyl-β-lactam derivative 3 at pH 7.5 and analyzed by the
MALDI mass spectrometry. Compound 3 was previously
used for the selective modification of the enamine-formable
nucleophilic lysine -amino group of aldolase antibodies (33).
At 30 min, 0-4 additions of 3 per molecule of FT-YLK3
were observed. At 2 h, 1-5 additions, and at 6 h 1-7
additions of 3 per molecule of FT-YLK3 were observed. FTYLK3 contains tyrosine, threonine, and cysteine residues that
can react with 3. Even with consideration of labeling at these
additional modification sites, the results indicated that more
than one amino group was labeled per peptide. Any of these
Small Designer Aldolase Enzymes
FIGURE 1: Enaminone formation of peptides FT-YLK3 (solid
circle), FT-YLK25 (square), and YLK-18opt (triangle) with 2,4pentanedione. The increase in absorption at 318 nm was monitored
in 5% CH3CN-42.5 mM sodium phosphate at 25 °C, [peptide]
100 µM, and [2,4-pentanedione] 5 mM. The relative velocity (Vrel)
was compared.
FIGURE 2: S-V plot of the FT-YLK-3-catalyzed retro-aldol reaction
of (()-4. Conditions: [FT-YLK3] 100 µM, 5% CH3CN-42.5 mM
sodium phosphate (pH 7.5), and 25 °C.
multiple amino groups could serve as a nucleophile for the
enamine-based catalysis. We could not determine the order
of the reactivity of the five lysyl residues of the peptide.
Retro-Aldol Reactions. Peptides FT-YLK3 and FT-YLK25
catalyzed the retro-aldol reactions of fluorogenic substrates
(()-4 and (()-anti-5 (Scheme 1). FT-YLK3- and FTYLK25-catalyzed reactions displayed saturation kinetics as
described by the Michaelis-Menten equation (Figure 2). The
kinetic parameters are shown in Table 2. The Km values were
in the millimolar range. The rate accelerations above the
buffer-catalyzed background reaction (kcat/kuncat) of peptides
FT-YLK3 and FT-YLK25 were 1400 and 1200 for (()-4
and 500 and 250 for (()-anti-5, respectively, and were 1.53-fold superior to the parental peptide YLK-18opt for the
both substrates, respectively.
When diketone 6 (2 mM) was added to a solution of the
peptide and then the retro-aldol reaction of 4 was performed,
the catalytic activity was inhibited. This result indicates that
the binding site for diketones and the catalytic site are shared.
Thus, together with the results of enaminone formation with
2,4-pentanedione as described above, these results indicate
that these peptides use an enamine mechanism for the retroaldol reactions.
For a family of aldolase antibodies, a correlation between
catalytic activity and reactivity to diketone was observed:
Aldolase antibody that catalyzed the retro-aldol reaction with
higher activity displayed higher reactivity to the corresponding diketone (24). A similar correlation was observed within
Biochemistry, Vol. 44, No. 20, 2005 7587
this family of aldolase peptides (peptides possessing aldolase
activity). Peptides FT-YLK3 and FT-YLK25 formed the
enaminone with 2,4-pentanedione 2.9- and 2.6-fold faster,
respectively, than YLK-18opt at pH 7.5 and catalyzed the
retro-aldol reaction of 4 2.6- and 2.2-fold, respectively, more
efficiently based on the kcat than YLK-18opt (Table 2).
Improvement in enaminone formation with the diketone
correlated with an improved kcat within this family of
peptides. Reaction-based selection with diketone provided
improved aldolase catalysts that operate via an enamine
mechanism.
Lysine -amino groups of the peptides should be essential
for catalysis. To examine the catalytic role of the lysine
residues, all five lysine residues in FT-YLK3 were replaced
by arginine residues. Peptide FT-YLK3-R5, with five arginines, displayed less than 5% of the catalytic activity of FTYLK3 when the reaction was performed using peptide (100
µM) and (()-4 (2 mM). This result was consistent with the
catalytic role for the -amino groups of the lysine residues
of FT-YLK3. To examine the effect of the N-terminal amine,
N-terminal acetylated peptide FT-YLK3-NAc was also
examined. Peptide FT-YLK3-NAc catalyzed the retro-aldol
reaction with the same velocity as that of FT-YLK3, when
the reaction was performed by using peptide (100 µM) and
(()-4 (2 mM). These results indicated that the N-terminal
amine of FT-YLK3 was not important for catalysis and lysine
-amino groups were essential for the catalysis.
MALDI mass analysis of FT-YLK3 revealed that the
peptide forms a covalent dimer, although a monomer was
the main species detected after storage in the reaction buffer
used for kinetic studies. The role of the cysteine residue that
formed the disulfide bond was assessed by synthesis of the
serine analogue FT-YLK3-23S. Peptide FT-YLK3-23S also
catalyzed the retro-aldol reaction of (()-4. The velocity of
the catalyzed reaction was 75% that of FT-YLK3 when the
reaction was performed using peptide (100 µM) and (()-4
(2 mM). Thus covalent dimer formation is not essential for
catalysis, but dimer formation enhanced the catalytic activity
and the folding (see below).
Peptides possessing lysine residue(s), but otherwise not
related to YLK-18opt, ELLELDKWASLWNC, ELKDKWASLWNWFNIT, ELDKWASLWNWFDITGGC, and CDEKSKLQEIYQELTQLKAAVGEL, were also analyzed for the
retro-aldol reaction of (()-4 under the same conditions. Their
activities were less than 5% that of FT-YLK3. In addition,
1 mM lysine, arginine, tyrosine, proline, mixtures of these
amino acids, or lysyllysine did not catalyze the retro-aldol
reactions of 4.
Studies performed using peptide (100 µM) and each of
the enantiomers of 4 (1 mM) in 10% CH3CN-40 mM
sodium phosphate (pH 7.5) revealed that the ratios of the
initial velocities of peptide-catalyzed reactions of (R)-4:(S)-4
were 1.3:1.0 for YLK-18opt, 1.8:1.0 for FT-YLK3, and 1.0:
1.0 for FT-YLK25. The selected terminal six residues of the
peptides thus had an effect on the enantioselectivity of the
peptide-catalyzed reactions. Although the selectivity was
moderate, these results indicate that the addition of amino
acid sequence can alter substrate specificity.
These peptides also have the potential to catalyze retroaldol reactions of natural sugars. When a reaction was
performed using FT-YLK3 (50 µM) and D-erythrose (5 mM)
in 42.5 mM sodium phosphate (pH 7.5), the initial velocity
7588 Biochemistry, Vol. 44, No. 20, 2005
Tanaka et al.
Table 2: Kinetic Parameters of Peptide-Catalyzed Retro-Aldol Reactions and Relative Velocity of Enaminone Formation
retro-aldol reaction of (()-4a
peptide
YLK-18opt
FT-YLK3
FT-YLK25
Km (mM)
kcat (min-1)
1.8
1.8
1.8
2.1 ×
5.6 × 10-4
4.7 × 10-4
10-4
kcat/kuncatc
540
1400
1200
retro-aldol reaction of (()-anti-5a
enamine formation with
2,4-pentanedioneb
Km (mM)
kcat (min-1)
kcat/kuncatc
Vrel
0.9
1.1
0.9
4.1 ×
1.2 × 10-3
5.9 × 10-4
170
500
250
1
2.9
2.6
10-4
a Reaction conditions: [peptide] 100 µM and 5% CH CN-42.5 mM sodium phosphate (pH 7.5) for (()-4 and 5% DMSO-42.5 mM sodium
3
phosphate (pH 7.5) for (()-anti-5, at 25 °C. b Conditions: [peptide] 100 µM, [2,4-pentanedione] 5 mM, and 5% CH3CN-42.5 mM sodium phosphate
(pH 7.5). c The first-order kinetic constant of the background reaction (kuncat) was 3.9 × 10-7 min-1 for (()-4 and 2.4 × 10-6 min-1 for (()-anti-5.
Scheme 2: (a) Peptide-Catalyzed Aldol Reactions and (b)
Peptide-Catalyzed Michael Reactions
FIGURE 3: Peptide-catalyzed aldol reactions using FT-YLK25
(square), FT-YLK3 (solid circle), YLK-18opt (triangle), and
background (open circle). (a) Reaction of acetone and 6-methoxy2-naphthaldehyde (7) to afford aldol 4. Conditions: [peptide] 100
µM, [acetone] 5% (680 mM), and [7] 2 mM in 5% CH3CN-40
mM sodium phosphate (pH 7.5). (b) Reaction of acetone and
4-nitrobenzaldehyde (8) to afford aldol 9. Conditions: [peptide]
100 µM, [acetone] 5% (680 mM), and [8] 2 mM in 5% CH3CN42 mM sodium phosphate (pH 7.0).
of the retro-aldol reaction was 3.7 × 10-3 µM/min after
background correction (background 5.4 × 10-4 µM/min).
Aldol Reactions. The peptides also catalyzed aldol reactions of acetone and aldehydes (Scheme 2a, Figure 3). The
reaction was performed by using peptide (100 µM) and
6-methoxy-2-naphthaldehyde (7) (2 mM) in 5% acetone (680
mM)-5% CH3CN-40 mM sodium phosphate (pH 7.5), and
the production of 4 was followed by HPLC assay. The initial
velocities of the formation of 4 were 4.6 × 10-2 µM/min
for FT-YLK25, 3.3 × 10-2 µM/min for FT-YLK3, and 2.2
× 10-2 µM/min for YLK-18opt after background correction
(background 2.3 × 10-3 µM/min) (Figure 3a). The velocities
of the peptide FT-YLK25- and FT-YLK3-catalyzed reactions
were 20- and 14-fold faster than that of the background and
were 2- and 1.5-fold faster than that of YLK-18opt,
respectively. The reactions with FT-YLK25 and FT-TLK3
afforded 173 and 110 µM 4 after 14 days, respectively. When
the reaction was performed by using peptide (100 µM) and
4-nitrobenzaldehyde (8) (2 mM) in 5% acetone (680 mM)5% CH3CN-40 mM sodium phosphate (pH 7.0) and the
production of 9 was followed by HPLC assay, the initial
velocities were 2.9 × 10-1 µM/min for FT-YLK25 and 2.0
× 10-1 µM/min for FT-YLK3 after background correction
(background 3.6 × 10-2 µM/min) (Figure 3b). The reactions
with FT-YLK25 and FT-TLK3 afforded 488 and 221 µM 9
after 22 h, respectively. Aldol products accumulated to a
higher concentration than the peptide catalyst concentration,
indicating that the peptides catalyzed the reaction with
turnover. The peptide-catalyzed reactions of aldehyde 8 (1
mM) and 1% acetone (136 mM) at pH 7.0 also provided the
corresponding aldol product 9. The initial velocity of the
FT-YLK3-catalyzed reaction was 5.4 × 10-2 µM/min
(background 3.0 × 10-3 µM/min), and the reaction afforded
190 µM 9 after 72 h. While FT-YLK3 was a slightly better
catalyst than FT-YLK25 for the retro-aldol reaction of 4,
peptide FT-YLK25 was a slightly better catalyst than FTYLK3 for the aldol reaction.
The enantioselectivity of these catalyzed reactions was
moderate: the FT-YLK3- and FT-YLK25-catalyzed reaction
afforded 10% ee and 8% ee of (R)-9, respectively, at 7 h
after background correction when the reactions were performed under the conditions described in Figure 3b. For FTYLK3, the (R)-configuration of the major product in the aldol
reaction was the same as that of the preferred substrate
stereochemistry in the retro-aldol reaction. This result was
similar to the aldolase antibody-catalyzed reactions: for
example, antibody 38C2 catalyzes the retro-aldol reaction
of (S)-4 and affords (S)-4 in the aldol reaction (34).
Michael Reaction. Since our peptides operate via an
enamine mechanism, they should also catalyze other reactions that exploit enamine intermediates. To examine the
catalytic breadth of these peptides, the Michael reaction of
acetone to maleimide 10 (35, 36) was performed (Scheme
2b). FT-YLK25 and serine mutant FT-YLK3-23S were used
for this reaction. When the reaction was performed using
peptide (100 µM) and 10 (1 mM) in 5% acetone (680 mM)2.5% CH3CN-2.5% DMSO-40 mM sodium phosphate (pH
7.0), the initial velocity of the formation of 11 was 7.6 ×
10-2 µM/min for FT-YLK25 and 6.4 × 10-2 µM/min for
FT-YLK3-23S (background 3.2 × 10-3 µM/min). The
velocities of the peptide-catalyzed reactions were approximately 20-fold faster than that of the background.
Small Designer Aldolase Enzymes
Biochemistry, Vol. 44, No. 20, 2005 7589
FIGURE 4: CD spectra of peptides: (a) YLK-18opt, (b) FT-YLK3, (c) FT-YLK3-23S, (d) FT-YLK25, (e) FT-YLK3-R25, and (f) FluS303FTYLK3. The CD spectra were recorded in 45 mM sodium phosphate (pH 7.5) at 25 °C at peptide concentrations of 100 µM, except for
FluS303-FTYLK3 which was recorded at 50 µM.
Relationship between Folding and Function. Since no
additional amino functionalities were selected in FT-YLK3
and FT-YLK25 compared to YLK-18opt, improvement of
the catalytic activity of the peptides may originate from
structural changes to the peptide and/or the addition of factors
such as acid/base catalysis and transition state stabilization.
To analyze the structures of the peptides, circular dichroism
(CD) studies were performed. The CD spectra of peptides
(100 µM) in 45 mM sodium phosphate buffer (pH 7.5) at
25 °C are shown in Figure 4. For FT-YLK3, the mean residue
ellipticities at 208 and 222 nm were -3.04 × 104 and -2.47
× 104 deg cm2 dmol-1, respectively, and the R-helical content
(37) was approximately 70% (the mean residue ellipticity
of FT-YLK3 was calculated as the monomer). FT-YLK25
also adopted an R-helical structure. The parent peptide YLK18opt showed a much reduced R-helical content (approximately 20%) under the same conditions. These results
suggest that the six amino acid residue extensions stabilize
the R-helical conformation of the peptides. The ratio of the
mean residue ellipticity of FT-YLK-3-23S at 208 and 222
nm was 0.77, similar to that of FT-YLK3 (0.81). The
R-helical content of FT-YLK3-23S was approximately 50%.
The retro-aldolase activity of these peptides was correlated
with R-helical content (Figure 5). These results indicate that
improved catalytic activity is intimately linked with stabilization of the conformation of the peptide. The conformation
of the peptides enables their lysine -amino groups to be
more nucleophilic than a lysine in an unstructured environment.
To understand whether the folding was related to selfassociation, the effect of peptide concentration on function
was analyzed. If the peptide functions as a monomer, a linear
relationship should be observed in the plot of peptide
concentration versus catalytic activity. If self-association of
the peptide affects folding and thus catalytic activity, the
FIGURE 5: Plot of relative velocity (Vrel) of the peptide-catalyzed
retro-aldol reaction of (()-4 versus R-helical content.
relationship should be nonliner. Since catalytic activity was
correlated to enaminone formation as described above, the
velocity of the enaminone formation with 2,4-pentanedione
was compared within a series of peptide concentrations. The
results showed a nonlinear relationship between the velocity
of enaminone formation and peptide concentration (Figure
6). Peptides showed an allosteric-type lag with a slower
velocity at a low peptide concentration, although peptides
FT-YLK3 and FT-YLK25 began to function at a lower
concentration than YLK-18opt. These results indicate that
the self-association of peptide molecules is clearly critical
to the function of the peptides. The self-association enhanced
the appropriate folding of the peptides and vice versa.
Addition of six amino acid residues to YLK-18opt improved
the folded state of the peptide and thus enhanced the selfassociation of the peptides and resulted in better function at
lower peptide concentrations.
Two possible mechanisms for the reactivity change of the
lysine -amino group in proteins have been reported: one
involves an electrostatic mechanism (a positively charged
residue within an interactive distance of the active site lysine
-amino group acts to modulate the pKa of a proximal amino
7590 Biochemistry, Vol. 44, No. 20, 2005
Tanaka et al.
Chart 1
FIGURE 6: Effect of peptide concentration on enaminone formation
with 2,4-pentanedione. Conditions: [2,4-pentanedione] 5 mM in
5% CH3CN-42.5 mM sodium phosphate (pH 7.5). Enaminone
formation was monitored at 318 nm. Key: FT-YLK3 (solid circle),
FT-YLK25 (square), FluS303-FTYLK3 (open circle), and YLK18opt (triangle).
group facilitating its reactivity) and the other is the creation
of hydrophobic microenvironments (36, 38, 39). Since the
lysine residues of the peptides are within amphipathic
R-helical structures and since the self-association and the
catalytic activity of the peptides are related, either mechanism
or a combination of the both mechanisms would be possible
for the peptides described here. It is important to note that
the peptides with improved activities and with improved
conformational stabilities required for catalysis have been
obtained by the reaction-based selection without complicated
structural design for the improvement of their catalytic
activities.
ImproVement of Substrate Specificity and Proficiency of
Peptide Catalysts. Although peptides FT-YLK3 and FTYLK25 showed better catalytic activity than peptide YLK18opt, the specificity of FT-YLK3 and of FT-YLK25 for
the substrates was moderate: their Km values were in the
millimolar range. To improve the substrate specificity, we
have developed a modular assembly strategy (40). This
strategy recruits a substrate-binding module to provide
enhanced substrate specificity in covalent combination of
binding and catalytic domain modules. When the binding
site is in close proximity to the catalytic site, the catalytic
site receives the benefit of a higher local substrate concentration provided by sequestering the substrate in close proximity
to the catalytic site. The potential advantages of this approach
are that it reduces the demand on the functionalization of
the catalytic site, which is limited in small peptides, and it
is modular, therefore making its adaptation to a variety of
specificities rapid.
To endow peptide FT-YLK3 with improved substrate
specificity, we sought to covalently connect this peptide to
another peptide that binds a substrate molecule with high
affinity, anticipating that the local concentration of substrate
proximal to the catalytic peptide would be increased. Rozinov
and Nolan reported small peptides that bind to small molecule
fluorophores (41). One of these peptides, YPNEFDWWDYYY (FluS303 in their paper), binds to fluorescein. This
peptide was selected from a combinatorial library of 12-mer
peptides displayed on phage. We chose this peptide as our
substrate-binding module and attached it to the N-terminus
of FT-YLK3. Since FT-YLK3 has a tyrosine residue at the
N-terminus and FluS303 has three tyrosine residues at the
C-terminus, one tyrosine was removed at the junction.
Peptide FluS303-FTYLK3 was chemically synthesized (Cterminal amide) and characterized.
Substrate 12, which contains a structural moiety from
fluorescein, was designed and used in the peptide-catalyzed
retro-aldol reactions. The kinetic parameters are shown in
Table 3. The Km value for the FluS303-FTYLK3-catalyzed
retro-aldol reaction of substrate 12 was 8 µM, while the Km
of the FluS303-FTYLK3-catalyzed retro-aldol reaction for
substrate 4 was 1.1 mM. The Km value of substrate 12 was
140-fold lower than that of 4 in the FluS303-FTYLK3catalyzed reaction, and the kcat/Km value of 12 was 43-fold
higher than that of 4. Thus FluS303-FTYLK3 successfully
discriminated between substrates 12 and 4. As calculated
from the Km values, the differential binding energy (∆∆G
) -RT ln{(Km of 12)/(Km of 4)} at 25 °C) in favor of
substrate 12 over 4 in the FluS303-FTYLK3-catalyzed
reactions was 2.9 kcal/mol. The catalytic module itself, FTYLK3, catalyzed the reaction of 12 with a Km of 130 µM.
Fortuitously, the Km value of substrate 12 was 14-fold lower
than that of substrate 4 in the FT-YLK3-catalyzed reactions.
Thus, for FT-YLK3-catalyzed reactions, differential binding
of substrate 12 over 4 was favored by 1.6 kcal/mol. Note
that the Km value of FluS303-FTYLK3 with substrate 12 was
16-fold lower than that of FT-YLK3 with 12 while the Km
values of 4 were similar in both the FluS303-FTYLK3- and
FT-YLK3-catalyzed reactions. These results indicate that the
enhanced substrate specificity of FluS303-FTYLK3 for 12
is the result of two effects. First, binding of 12 to the FTYLK3 domain as compared to substrate 4 is enhanced by
1.6 kcal/mol. While the exact mechanism for this increase
is not clear at present and was not part of the design strategy,
heteroatom presentation within 12 may play a role. Second,
the addition of the FluS303 sequence to FTYLK3 further
increased the affinity of the peptide for substrate 12 (∆∆G
) 1.7 kcal/mol) (∆∆G ) -RT ln{(Km of FluS303-FTYLK3catalyzed reaction of 12)/(Km of FT-YLK3-catalyzed reaction
of 12)} at 25 °C) while having a minimal effect on the
binding of substrate 4 (∆∆G ) 0.3 kcal/mol) (∆∆G ) -RT
ln{(Km of FluS303-FTYLK3-catalyzed reaction of 4)/(Km of
FT-YLK3-catalyzed reaction of 4)} at 25 °C). We assign
this increase in substrate specificity to increase in local
substrate concentration at the catalytic site due to the effect
of the modular assembly strategy.
Further, when the FluS303-FTYLK3-catalyzed reaction of
12 was performed in the presence of fluorescein (500 µM),
the kinetic parameters were Km 60 µM, kcat 2.5 × 10-4 min-1,
and kcat/Km 4.2 M-1min-1. Thus, the Km value substantially
increased while the kcat remained unchanged (within experimental deviation). This result indicates that when fluorescein
occupies the substrate-binding domain, in competition with
12, the specificity constant is greatly reduced. This experiment also validated the modular assembly strategy.
Although the kcat values of peptide catalysts were lower
than those of much larger designer protein catalysts such as
Small Designer Aldolase Enzymes
Biochemistry, Vol. 44, No. 20, 2005 7591
Table 3: Kinetic Parameters of Peptide-Catalyzed Retro-Aldol Reactionsa
peptide
substrate
Km (µM)
kcat (min-1)
kcat/Km
(M-1 min-1)
kcat/kuncat
(kcat/Km)/kuncat
(M-1)
(kcat/Km)/kuncat
per residue (M-1)
FluS303-FTYLK3
FT-YLK3
FluS303-FTYLK3
FT-YLK3b
12
12
4
4
8
130
1100
1800
2.3 × 10-4
2.0 × 10-4
7.4 × 10-4
5.6 × 10-4
29
1.5
0.67
0.31
1800
1500
1900
1400
2.2 × 108
1.2 × 107
1.7 × 106
8.0 × 105
6.3 × 106
4.9 × 105
4.9 × 104
3.3 × 104
a Reaction conditions: [peptide] 50 µM except as noted and 10% CH CN-42.5 mM sodium phosphate (pH 7.5) for substrate 12 and 5% CH CN3
3
42.5 mM sodium phosphate (pH 7.5) for substrate 4 at 25 °C. b [Peptide] 100 µM. See Table 2.
FIGURE 7: Comparison of designer peptide catalysts and antibody
catalysts. Plot of (kcat/Km)/kuncat per residue versus number of amino
acid residues found in the peptides and antibody catalysts. Peptides
are YLK-opt18, FT-YLK3, FT-YLK25, and FluS303-FTYLK3.
For aldolase antibodies IgG38C2, IgG33F12, and Fab28, see ref
24. Key: substrate 4 (solid squares) and substrate 12 (diamonds).
aldolase antibodies, the catalytic proficiency (42) per residue
((kcat/Km)/kuncat per residue) of the FluS303-FTYLK3catalyzed reaction of 12 was in the range of those of larger
protein catalysts (Figure 7). Peptide FluS303-FTYLK3,
which consists of only 35 amino acid residues, had a catalytic
proficiency per residue of 6 × 106 M-1 (Table 3). This value
was comparable to those of antibody catalysts (approximately
1500 and 500 amino acid residues as IgG and Fab form,
respectively) on a per residue basis. Some natural enzymes
also have catalytic proficiency per residue of approximately
106 M-1 (42). Thus, the 35-mer peptide FluS303-FTYLK3
functioned at near the catalytic proficiency per residue
observed for some larger protein catalysts.
FluS303-FTYLK3 also formed a UV-observable enaminone upon addition of 2,4-pentanedione. As indicated in
Figure 6, FluS303-FTYLK3 also showed a nonlinear relationship between the enaminone formation velocity and
peptide concentration. The concentration necessary to function was approximately 15 µM for FluS303-FTYLK3, similar
to that of FT-YLK3 (approximately 5 µM). Appending the
substrate-binding domain to FTYLK3 did not affect the
function of the catalytic domain. The CD spectra of FluS303FTYLK3 (50 µM) in 45 mM sodium phosphate (pH 7.5) at
25 °C indicated that the peptide contained R-helical and
β-sheet structures (Figure 4f). Studies using the program k2d
(43) indicate the R-helical content was 27% (maximum error
22.7%). Since the R-helical content of FT-YLK3 (100 µM)
was estimated to be approximately 70%, the ideal R-helical
content of FluS303-FTYLK3 was calculated to be 48% (24
residues × 0.7/35 residues). Denaturation of the R-helical
structure (4.5 M guanidine hydrochloride) as observed in the
CD spectra was accompanied by complete loss of catalytic
activity. Therefore, with peptide FluS303-FTYLK3, the
R-helical structure also correlated with the function.
We have developed small peptides that catalyze carboncarbon bond formation and cleavage via an enamine mech-
anism. The reaction-based selection with designed compounds afforded improved small peptide catalysts with
respect to both catalytic activity and folded state. We have
demonstrated that the reaction-based selection using phage
display libraries, which has been effectively used to generate
larger protein catalysts (23, 24, 44), is useful for the
development of small designer peptide catalysts. We have
also demonstrated that the substrate specificity of the peptide
catalyst for a small molecule substrate can be improved by
a modular assembly strategy, i.e., covalent conjugation of a
binding domain to the catalytic peptide module. We conclude
that the combination of design and reaction-based selection
from libraries effectively provides small designer enzymes
with good rate acceleration and excellent substrate specificity.
REFERENCES
1. Reetz, M. T. (2004) Controlling the enantioselectivity of enzymes
by directed evolution: practical and theoretical ramifications, Proc.
Natl. Acad. Sci. U.S.A. 101, 5716-5722.
2. Joerger, A. C., Mayer, S., and Fersht, A. R. (2003) Mimicking
natural evolution in vitro: An N-acetylneuraminate lyase mutant
with an increased dihydrodipicolinate synthase activity, Proc. Natl.
Acad. Sci. U.S.A. 100, 5694-5699.
3. Seebeck, F. P., and Hilvert, D. (2003) Conversion of a PLPdependent racemase into an aldolase by a single active site
mutation, J. Am. Chem. Soc. 125, 10158-10159.
4. Dwyer, M. A., Looger, L. L., and Hellinga, H. W. (2004)
Computational design of a biologically active enzyme, Science
304, 196-197.
5. Xu, Y., Yamamoto, N., and Janda, K. (2004) Catalytic antibodies: hapten design strategies and screening methods, Bioorg. Med.
Chem. 12, 5247-5268.
6. Tanaka F. (2002) Catalytic antibodies as designer proteases and
esterases, Chem. ReV. 102, 4885-4906.
7. Tanaka, F., and Barbas, C. F., III (2002) Reactive immunization:
a unique approach to catalytic antibodies, J. Immunol. Methods
269, 67-79.
8. Johnsson, K., Allemann, R. K., Widmer, H., and Benner, S. A.
(1993) Synthesis, structure and activity of artificial, rationally
designed catalytic polypeptides, Nature 365, 530-532.
9. Perez-Paya, E., Houghten, R. A., and Blondell, S. E. (1996)
Functionalized protein-like structures from conformationlly defined
synthetic combinatorial libraries, J. Biol. Chem. 271, 4120-4126.
10. Allert, M., Kjellstrand, M., Broo, K., Nilsson, A., and Baltzer, L.
(1998) A designed folded polypeptide model system that catalyzes
the decarboxylation of oxaloacetate, J. Chem. Soc., Perkin Trans.
2, 2271-2274.
11. Allert, M., and Baltzer, L. (2002) Setting the stage for new catalytic
functions in designed proteinssExploring the imine pathway in
the efficient decarboxylation of oxaloacetate by an Arg ( Lys
site in a four-helix bundle protein scaffold, Chem. Eur. J. 8, 25492560.
12. Weston, C. J., Cureton, C. H., Calvert, M. J., Smart, O. S., and
Allemann, R. K. (2004) A stable miniature protein with oxaloacetate decarboxylase activity, ChemBioChem 5, 1075-1080.
13. Broo, K., Brive, L., Ahlberg, P., and Baltzer, L. (1997) Catalysis
of hydrolysis and transesterification reactions of p-nitrophenyl
esters by a designed helix-loop-helix dimer, J. Am. Chem. Soc.
119, 11362-11372.
14. Nilsson, J., and Baltzer, L. (2000) Reactive-site design in foldedpolypeptide catalystssthe leaving group pKa of reactive esters sets
7592 Biochemistry, Vol. 44, No. 20, 2005
the stage for cooperativity in nucleophilic and general-acid
catalysis, Chem. Eur. J. 6, 2214-2220.
15. Douat-Casassus, C., Darbre, T., and Reymond, J.-L. (2004)
Selective catalysis with peptide dendrimers, J. Am. Chem. Soc.
126, 7817-7816.
16. Wei, Y., and Hecht, M. H. (2004) Enzyme-like proteins from an
unselected library of designed amino acid sequences, Protein Eng.
Des. Sel. 17, 67-75.
17. Rutledge, S. E., Volkman, H. M., and Schepartz, A. (2003)
Molecular recognition of protein surfaces: high affinity ligands
for the CBP KIX domain, J. Am. Chem. Soc. 125, 14336-14336.
18. Smith, G. D., Eisenthal, R., and Harrison, R. (1977) Determination
of dissociation and Michaelis constants at near-equal enzymesubstrate concentrations, Anal. Biochem. 79, 643-647.
19. Heine, A., DeSantis, G., Luz, J. G., Mitchell, M., Wong, C. H.,
and Wilson, I. A. (2001) Observation of covalent intermediates
in an enzyme mechanism at atomic resolution, Science 294, 369374.
20. Allard, J., Grochulski, P., and Sygusch, J. (2001) Covalent
intermediate trapped in 2-keto-3-deoxy-6-phosphogluconate (KDPG)
aldolase structure at 1.95-Å resolution, Proc. Natl. Acad. Sci.
U.S.A. 98, 3679-3684.
21. Choi, K. H., Shi, J., Hopkins, C. E., Tolan, D. R., and Allen, K.
N. (2001) Snapshots of catalysis: the structure of fructose-1,6(bis)phosphate aldolase covalently bound to the substrate dihydroxyacetone phosphate, Biochemistry 40, 13868-13875.
22. Littlechild, J. A., and Watson, H. C. (1993) A data-based reaction
mechanism for type I fructose bisphosphate aldolase, Trends
Biochem. Sci. 18, 36-39.
23. Tanaka, F., Lerner, R. A., and Barbas, C. F., III (2000) Reconstructing aldolase antibodies to alter their substrate specificity and
turnover, J. Am. Chem. Soc. 122, 4835-4836.
24. Tanaka, F., Fuller, R., Shim, H., Lerner, R. A., and Barbas, C. F.,
III (2004) Evolution of aldolase antibodies in vitro: correlation
of catalytic activity and reaction-based selection, J. Mol. Biol. 335,
1007-1018.
25. Gilbert, H. F., III, and O’Leary, M. H. (1975) Modification of
arginine and lysine in proteins, Biochemistry 14, 5194-5199.
26. Otwell, H. B., Cipollo, K. L., and Dunlap, R. B. (1979)
Modification of lysyl residues of dihydrofolate reductase with 2,4pentanedione, Biochim. Biophys. Acta 568, 297-306.
27. Wagner, J., Lerner, R. A., and Barbas, C. F., III (1995) Efficient
aldolase catalytic antibodies that use the enamine mechanism of
natural enzymes. Science 270, 1797-1800.
28. Tanaka, F., and Barbas, C. F., III (2001) Phage display selection
of peptides possessing aldolase activity, Chem. Commun., 769770.
29. Vlahos, C, J., and Dekker, E. E. (1986) Amino acid sequence of
the pyruvate and the glyoxylate active-site lysine peptide of
Escherichia coli 2-keto-4-hydroxyglutarate aldolase, J. Biol. Chem.
261, 11049-11055.
30. Bjornestedt, R., Zhong, G., Lerner, R. A., and Barbas, C. F., III
(1996) Copying nature’s mechanism for the decarboxylation of
Tanaka et al.
β-keto acid into catalytic antibodies by reactive immunization, J.
Am. Chem. Soc. 118, 11720-11724.
31. Barbas, C. F., III, Burton, D. R., Scott, J. K., and Silverman, G.
J., Eds. (2001) Phage Display: A Laboratory Manual, Cold Spring
Harbor Laboratory Press, Cold Spring Harbor, NY.
32. Lide, D. R., Ed. (2002) CRC Handbook of Chemistry and Physics,
83rd ed., p 7-1, CRC Press, New York.
33. Tanaka, F., Lerner, R. A., and Barbas, C. F., III (1999) Thiazoliumdependent catalytic antibodies produced using a covalent modification strategy, Chem. Commun., 1383-1384.
34. Zhu, X., Tanaka, F., Hu, Y., Heine, A., Fuller, R., Zhong, G.,
Olson, A. J., Lerner, R. A., Barbas, C. F., III, and Wilson, I. A.
(2004) The origin of enantioselectivity in aldolase antibodies:
crystal structure, site-directed mutagenesis, and computational
analysis, J. Mol. Biol. 343, 1269-1280.
35. Tanaka, F., Thayumanavan, R., and Barbas, C. F., III (2003)
Fluorescent detection of carbon-carbon bond formation, J. Am.
Chem. Soc. 125, 8523-8528.
36. Tanaka, F., Thayumanavan, R., Mase, N., and Barbas, C. F., III
(2004) Rapid analysis of solvent effects on enamine formation
by fluorescence: how might enzymes facilitate enamine chemistry
with primary amines?, Tetrahedron Lett. 45, 325-328.
37. Korsgren, P., Ahlberg, P., and Baltzer, L. (2000) Covalent control
of polypeptide folding. Induction of helix-loop-helix motifs by
bridging, J. Chem. Soc., Perkin Trans. 2, 643-647.
38. Barbas, C. F., III, Heine, A., Zhong, G., Hoffmann, T., Gramatikova, S., Bjornestsdt, R., List, B., Anderson, J., Stura, E. A.,
Wilson, I. A., and Lerner, R. A. (1997) Immune versus natural
selection: Antibody aldolases with enzymic rates but broader
scope, Science 278, 2085.
39. Westheimer, F. H. (1995) Coincidences, decarboxylation, and
electrostatic effects, Tetrahedron 51, 3.
40. Tanaka, F., and Barbas, C. F., III (2002) A modular assembly
strategy for improving the substrate specificity of small catalytic
peptides, J. Am. Chem. Soc. 124, 3510-3511.
41. Rozinov, M. N., and Nolan, G. P. (1998) Evolution of peptides
that modulate the spectral qualities of bound, small-molecule
fluorophores, Chem. Biol. 5, 713-728.
42. Radzicka, A., and Wolfenden, R. (1995) A proficient enzyme,
Science 267, 90-93.
43. Andrate, N., Chacon, P., Merelo, J. J., and Moran, F. (1993)
Evolution of secondary structure of proteins from UV circular
dichroism spectra using an unsupervised learning neural network,
Protein Eng. 6, 383-390.
44. Cesaro-Tadic, S., Lagos, D., Honegger, A., Rickard, J. H.,
Partridge, L. J., Blackburn, G. M., and Pluckthun, A. (2003)
Turnover-based in vitro selection and evolution of biocatalysts
from a fully synthetic antibody library, Nat. Biotechnol. 21,
679-685.
BI050216J